The core soil of an embankment dam can be exposed to deteriorating processes, i.e., different kinds of internal erosion due to high hydraulic gradients, disadvantageous particle size distribu-tion, too coarse-grained filters or built-in defects. During internal erosion, fines from the core soil are washed out by the seepage, decreasing the impervious properties of the core. If the internal erosion process is discovered in time, drilling and grouting can be performed to stop the erosion. During drilling and grouting, eroded material from the core soil is replaced.

In this paper, the methodology: “Identification – Localization – Characterization – Remediation” has been proposed. The methodology was tested on a large-scale embankment dam in a laboratory environment. The dam had a central core of moraine and was built inside a watertight concrete structure so a reservoir of water could be created upstream the dam. The left abutment of the dam had higher seepage rates than the rest of the dam and therefore had to be remediated.

During the identification and localization phase, a 10 x 10 cm horizontal, high hydraulic conduc-tivity zone through the core soil was identified and localized at the left abutment at 1 m depth. During drilling at the abutment, it was found that the core soil beneath the damage had become more wet compared to when built. The remedial method used was compaction grouting with a new developed type of non-hardening grout material. The grouting pressures equaled the height of the vertical grout material column with an additional pressure of ~50 kPa to compensate for frictional losses during injection. The grout material was delivered via a novel pipe system where water and air were allowed to be drained. The seepage was lowered by 44 % directly after grout-ing and 60% four months after grouting.

Modelling can be used as a tool for prediction of the behaviour of embankment dams as a part of the dam safety work. It is advantageous to predict the performance and compare to measurements done, to obtain more knowledge about the dam behaviour, as these structures are complex and potential failures are hazardous. The research presented in this thesis covers parameter identification by backanalysis, interpretation of dam measurements and numerical predictions of dam behaviour. The research highlights the role of numerical modelling as a supportive tool in dam engineering, ratherthan a standalone technique. Two embankment dams were analysed in the research: a 45 metres high existing hydropower dam and a four metres high experimental dam built during the project.

The soil materials in an embankment dam vary significantly, as the zones in a dam have different functions. To create reliable numerical models, parameter values defining the stress-strain relationship of the materials are needed. Obtaining such information for existing embankment dams poses challenges, often due to limited available data and the potential risks associated with traditional field sampling methods. In previous research at Luleå University of Technology, inverse analysis was successfully applied to embankment dam calibration of finite element models against field measurements, by utilizing an optimisation code with a genetic algorithm for optimisation. Inverse analysis provides a non-destructive method for obtaining information about the stress-strain relationship of the material in a dam.

First, applications of inverse analysis are exemplified on an existing embankment dam. The study investigates the impact on the inverse analysis methodology when errors occur in the field measurements. The employed genetic algorithm showed its robustness when dealing with errors, this is important since errors are likely to occur in field measurements. Thereafter, the study examined the usage of parameters identified through inverse analysis in predictions of deformations when a stabilising berm was constructed on the downstream shoulder. The predicted deformations were compared to deformations from inclinometer measurements. The trend of the measured deformations was replicated in the numerical model, and the magnitudes were in the right order. The study shows that predicting future dam behaviour based on results from inverse analysis can be done reasonably well in this case.

Second, the mechanical behaviour of an experimental embankment dam is interpreted and modelled. Monitoring of pore pressure was done with transducers that were installed at different levels covering the whole core and parts of the filters. Measurements were performed continuously. The response of pore pressure in the core, during impoundment and operation, are focused on. A significant delay of the saturation front was observed, as the pore pressure in the bottom of the downstream part of the core was not building up as expected during impoundment and operation. Fully-coupled numerical analyses were performed, in order to better understand the conditions of the core in the experimental dam. The core was initially assumed to be homogeneous, but the numerical results showed poor agreement with the observed behaviour from field. By further analysing the measurements and modelling, the experimental dam was found to be non-homogeneous, even though it was built under very controlled conditions. Variations in the hydraulic conductivity in the dam core were therefore introduced in the numerical model. The hydraulic conductivity changed with height in the dam, was different in the vertical and horizontal direction and was also changing with time at specific places in the core. With these numerical adjustments better correlations against measurements were obtained, compared to the homogeneous case, indicating that homogeneous conditions are not suitable for the core. The study also showed that the values of parameters obtained from laboratory testing are not suitable for the whole core, as the conditions assumed in laboratory do not correspond to the prevailing field conditions.

Measurements of the strain development in the bottom of the embankment dam was done by fibre optics. The settlements in the dam body after construction are captured, mainly as areas of higher lateral strain at the toes of the dam. The 1% sloping foundation towards the downstream side is captured. The measurements show that more shear is activated at the downstream side. Impoundment causes the largest strains, it can be observed that the strains vary in the different dam zones the bottom of the dam body.

In summary, the research presented in this thesis has shown how numerical modelling can be used as support in dam engineering, when combined with monitoring data. Values of parameters that would have been difficult to retrieve otherwise are obtained by inverse analysis, making it possible to perform more reliable predictions. The modelling has also helped to explain unexpected behaviour from monitoring of pore pressure. When the experimental dam was built, it was expected that the core would be homogeneous. The monitoring of the dam and the numerical modelling revealed that the core was non-homogeneous. The experimental dam is small, and it was constructed under very controlled forms. Therefore, it is reasonable to assume that it would be difficult to construct a homogeneous core in a real, large embankment dams. This is an important finding in the thesis, which can influence both how dams should be numerically modelled as well as how dam safety assessments during first impoundment and the beginning of the operation phase should be done.

Numerical modelling is often needed as a tool to predict the behaviour and assess the safety of dam structures. Embankment dam structures analyses are quite complex and potential failures are hazardous. Predictions of dam behaviour by numerical modelling rely on knowledge about the mechanical properties of the materials the dam is constructed with. The materials included in a dam vary significantly because zones in the dam have different functions. In order to conduct reliable modelling, parameter values defining the stress-strain relationship of the materials are needed to be assigned.

 Obtaining information about the mechanical behaviour in already existing embankment dams is usually challenging. As many dams are old, there might be a limited amount of information available of the materials used, construction methods and mostly about the stress-strain relationship of the soil. Traditionally, field sampling is performed in order to obtain such information. However, conventional field sampling might negatively affect the dam body and thereby the performance as well as the safety of the dam. This is of special importance if sampling is performed in the impervious (core) part. Since traditional sampling might harm the dam body, use of non-destructive methods would be advantageous to utilise for obtaining information about the stress-strain relationship and the strength in a dam structure.

 An option for a non-destructive method is parameter identification by inverse analysis. The idea of inverse analysis is to calibrate finite element models towards field measurements. In the calibration process, the input for a stress-strain relationship (constitutive model) is modified until the discrepancy between the output of the numerical model and the associated chosen field measurement is minimised. The agreement between output from the numerical model and reality is measured by an objective function that will calculate the error. In order to automatically search for the minimum a search algorithm is utilised in the optimisation process. When the objective function is minimised, the calibration of the material parameters is done.

 In previous research at Luleå University of Technology, the method of inverse analysis was applied to an embankment dam. The finite element program PLAXIS was used in combination with an optimisation code. The optimisation code includes an objective function (for error evaluation) and a search algorithm. The genetic algorithm was employed as search algorithm, since it is known for its robustness and efficiency as well as the fact that it provides a set of solutions instead of one unique answer. This is beneficial from a geotechnical point of view, since engineering judgement can be included in the final choice of solution.   

 The first study in the present thesis deals with a case study of an embankment dam, where a simple model calibration was performed. This was a part of a larger study, at the ICOLD Benchmark Workshop in 2017, where the work presented here was forming one of the contributions. In order to have a model response similar to reality, the contributors were asked to choose constitutive models and calibrate them. The calibration was done by manually changing the input for the constitutive model chosen. While the response of the numerical finite element model was capturing the trends of measured total stresses and pore pressure in the dam quite well, there were difficulties in capturing the long term deformations of the dam. This was a challenge for all contributors. An idea for improving the model response, is to run a more advanced calibration by inverse analysis.

 In the second study in the thesis, predictions are presented for the embankment dam that inverse analysis was previously conducted for at LTU. Strengthening actions in form of a new berm were performed at the dam. With identified material parameter values from the inverse analysis, predictions were conducted both before and after the strengthening measures. The predicted deformations were compared to deformation data from inclinometer measurements.  A reasonably well agreement was obtained with the real deformations. The trend of the deformations was replicated and the magnitudes of the deformations were in the right order. The study is indicating that predicting future dam behaviour based on results from inverse analysis can be done reasonably well.

 In the third and final study in the thesis, effects of random measurement error on the performance of the genetic algorithm for soil parameter identification are assessed. Also here, with the application to the embankment dam used in previous research at LTU. Optimisations were performed against inclinometer measurements. To be sure that the constitutive model can find the correct solution, synthetic (i.e. numerically generated) inclinometer data was utilised. Perturbations were randomly generated within chosen intervals of error and added to the numerically generated deformations.

 The genetic algorithm showed its robustness, by continuing to search for solutions without breaking down even if the field data was substantially perturbed. Considering usual errors for inclinometer measurements, the genetic algorithm can deliver good solutions. The inclinometer errors used were taken from literature, and thereafter related to the perturbations of the numerically generated data. Dealing with errors that are becoming gradually larger than what can be considered as usual, problems are faced by the genetic algorithm. In this cases it is difficult to find a solution, and if solutions are found they might significantly deviate from the unperturbed optimum solution.

 The three studies handled in this thesis are treating aspects of back analysis of embankment dams; from a simple calibration, to predictions based on material parameters from advanced inverse analysis and finally effects of errors on the genetic algorithm. It been shown that using inverse analysis for already existing embankment dams is very beneficial for the material characterisation and is forming a step towards better predictions of future dam behaviour.

Fyllningsdammen i Burvattnet uppfördes på 1940-talet av jordmaterial som fanns på platsen. I samband med ombyggnation av bottenutskovet skulle fyllningsdammen rivas och återuppbyggas, vilket innebar en möjlighet att studera och dokumentera effekter av åldrande av dammen.

Field sampling for evaluation of mechanical behaviour in embankment dams is not easily performed, because the performance and the safety of the structure may be unfavourably affected. A non-destructive method, inverse analysis, is an alternative. In this study, inverse analysis has been utilised to identify values of soil parameters for an embankment dam. An objective function and a genetic search algorithm were combined with finite-element software to perform the analysis. Values of model parameters were calibrated until inclinometer deformations from monitoring and computations corresponded to each other. Errors in field measurements occur, related to – for example – measurement precision, as well as handling and installation of the equipment. Search algorithms in mathematical optimisation might incur numerical problems if they are used against data containing errors. The performance of the genetic algorithm was investigated for the dam studied, when identification was performed against inclinometer data containing known errors of different magnitudes. The results showed that the genetic algorithm can search for solutions without obtaining numerical problems, even though the field data are substantially perturbed. It was found that the genetic algorithm is able to find good solutions for data from field measurements, including the usual errors in practical dam applications.

Over the lifetime of a dam several measures are usually taken in order to assure the stability and the performance of the dam. In this case a hydropower dam in Northern Sweden is in need of dam safety measures. The question arose, what consequences there might be when such measures are performed. In order to estimate these effects, simulations have been carried out in the finite element programme PLAXIS 2D. Thereby, the deformations and the stability of the dam for the planned work can be evaluated. The performed simulations are based upon previously conducted research at Luleå University of Technology, where soil parameters in the investigated dam were identified by a method of inverse analysis.Three sections have been analysed: A, B and C. In section A increasing pore water pressure has been observed at the downstream side of the dam. Thereby it has been concluded that a new drainage system is needed; new trenches of large size are to be excavated. In section B new toe berms are planned, due to the requirement that the dam should be able to divert leakage without erosion occurring at the dam toe. This contains soil material that might degrade when stresses are increased, with intensified deformations as a consequence. In section C a new berm is to be constructed, before this can be conducted an excavation is performed at the toe of the dam.The results have shown deformations of an acceptable magnitude and factors of safety that indicate conditions for the planned dam safety measures. Numerical values of deformations and factors of safety can be utilised as an attempt to establish alarm values for the stability of the dam. The finite element method is a useful tool for this kind of evaluation.

It is usually difficult to determine values for soil parameter values in earth- A nd rockfill dams by traditional methods. Field sampling is not easily performed, especially in the impervious parts, since the performance and safety of the dam structure may be affected in an unfavourable way. Therefore other methods, preferably non-destructive, are needed to investigate the mechanical behaviour. Inverse analysis has been utilised to identify soil parameter values for an earth- A nd rockfill dam. An error function and a genetic search algorithm were combined with a finite element software to perform the analysis. The model parameters in the chosen constitutive model were calibrated until the horizontal deformations corresponded to the horizontal inclinometer deformations. Errors or irregularities in field measurements can occur, for instance based on the accuracy of the equipment. In this study, the performance of the genetic algorithm was investigated, when applied to identify soil parameters for a dam. Added perturbations to simulated inclinometer data are randomly generated within a chosen interval of error. The results showed that the genetic algorithm found a minimum for the error function even though the field data was substantially perturbed. Errors up to 10% were shown to have minor impact

When performing predictions of future behaviour and assessments of the safety for embankment dams, numerical modelling if often needed as support. For embankment dams a usual issue is to obtain values of the material parameters for the material in the dam. Sampling is not easily performed, and it could also negatively affect the performance and safety of the dam. One way to determine values of the material parameters is to create a digital copy of the dam by utilizing so called inverse analysis. The created numerical model is calibrated towards field measurements. To make the reality and the numerical model correspond to each other, the values of material parameters for constitutive models are calibrated. The calibration is done by an automatized process. Further, the calibrated model could be used in predictions of future dam behaviour. The methodology has been shown to work for field data containing errors of a usual magnitude.